Dear Zelig ListServ Members,
There seems to be a bug in Zelig's mlogit command when
used under the latest version of R (2.4.0).
I have re-installed Zelig, and I still have the same
problem.
I am including a copy of my attempt to run one of the
examples you have included with the Zelig
documentation, and the resulting error:
> data(mexico)
>
> z.out1 <- zelig(as.factor(vote88) ~ pristr + othcok
+ othsocok,
+ model = "mlogit", data = mexico)
[I have removed some standard objects attached and
masked output, but I can send them to you if you need
them.]
Error in a[[n + 1]] <- value : object is not
subsettable
There does not seem to be a problem with running the
program under 2.3. Also, the mlogit.bayes function
seems to run fine.
z.out1 <- zelig(as.factor(vote88) ~ pristr + othcok +
othsocok,
+ model = "mlogit.bayes", data =
mexico)
Calculating MLEs and large sample var-cov matrix.
This may take a moment...
>
> summary(z.out1)
I am running R under Windows XP, and am using the
SciViews console.
Thanks in advance for you attention to this matter.
Sincerely,
Milton Mayfield
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/
I think questions similar to this have been posted previously, but I am
trying to determine if zelig will calculate clustered standard errors
for glm model.
This code
z.out<-zelig(y~x1+x2....,model="logit",robust=TRUE,cluster=x3,data=data)
returns this error:
Error in glm.control(...) : unused argument(s) (cluster ...)
are clustered standard errors possible with a logit model? The only
reference I can find for them in the documentation is for exponential
duration regression models.
thanks,
Ryan
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=zelig
Zelig program information: http://gking.harvard.edu/zelig/